flan-t5-xl-deepspeed-zero3-summary
This model is a fine-tuned version of google/flan-t5-large on the cnn_dailymail dataset. It achieves the following results on the evaluation set:
- Loss: 1.4054
- Rouge1: 42.6105
- Rouge2: 20.2181
- Rougel: 29.7866
- Rougelsum: 39.4431
- Gen Len: 98.0013
Model description
More information needed
Intended uses & limitations
More information needed
Training and evaluation data
More information needed
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.0001
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- distributed_type: multi-GPU
- num_devices: 10
- total_train_batch_size: 80
- total_eval_batch_size: 80
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 1
Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len |
|---|---|---|---|---|---|---|---|---|
| 1.6011 | 1.0 | 3589 | 1.4054 | 42.6105 | 20.2181 | 29.7866 | 39.4431 | 98.0013 |
Framework versions
- Transformers 4.26.0
- Pytorch 1.12.0
- Datasets 2.9.0
- Tokenizers 0.13.3